Articles

Health of men, women, and children in post-traffi cking services in Cambodia, Thailand, and Vietnam: an observational cross-sectional study

Ligia Kiss, Nicola S Pocock, Varaporn Naisanguansri, Soksreymom Suos, Brett Dickson, Doan Thuy, Jobst Koehler, Kittiphan Sirisup, Nisakorn Pongrungsee, Van Anh Nguyen, Rosilyne Borland, Poonam Dhavan, Cathy Zimmerman

Summary Background Traffi cking is a crime of global proportions involving extreme forms of exploitation and abuse. Yet little Lancet Glob Health 2015; research has been done of the health risks and morbidity patterns for men, women, and children traffi cked for various 3: e154–61 forms of forced labour. See Comment page e118 London School of Hygiene & Methods We carried out face-to-face interviews with a consecutive sample of individuals entering 15 post-traffi cking Tropical Medicine, London, UK (L Kiss PhD, C Zimmerman PhD, services in Cambodia, Thailand, and Vietnam. We asked participants about living and working conditions, experience N S Pocock MSc); International of violence, and health outcomes. We measured symptoms of anxiety and depression with the Hopkins Symptoms Organization for Migration, Checklist and post-traumatic stress disorder with the Harvard Trauma Questionnaire, and used adjusted logistic Bangkok, Thailand (D Thuy MA, regression models to estimate the eff ect of traffi cking on these mental health outcomes, controlling for age, sector of V A Nguyen MA, B Dickson BA, P Dhavan MPH, R Borland MA, exploitation, and time in traffi cking. N Pongrungsee, K Sirisup); International Organization for Findings We interviewed 1102 people, of whom 1015 reached work destinations. Participants worked in various sectors Migration, Phnom Penh, including sex work (329 [32%]), fi shing (275 [27%]), and factories (136 [13%]). 481 (48%) of 1015 experienced physical Cambodia (S Suos MA); and International Organization for violence, sexual violence, or both, with 198 (35%) of 566 women and girls reporting sexual violence. 478 (47%) of Migration, Hainoi, Vietnam 1015 participants were threatened and 198 (20%) were locked in a room. 685 (70%) of 985 who had data available (J Koehler, V Naisanguansri) worked 7 days per week and 296 (30%) of 989 worked at least 11 hours per day. 222 (22%) of 983 had a serious injury Correspondence to: at work. 61·2% (95% CI 58·2–64·2) of participants reported symptom of depression, 42·8% (39·8–45·9) reported Dr Ligia Kiss, London School of symptoms of anxiety, and 38·9% (36·0–42·0) reported symptoms of post-traumatic stress disorder. 5·2% (4·0–6·8) Hygiene & Tropical Medicine, Faculty of Public Health, had attempted suicide in the past month. Participants who experienced extremely excessive overtime at work, Department of Global Health restricted freedom, bad living conditions, threats, or severe violence were more likely to report symptoms of and Development, depression, anxiety, and post-traumatic stress disorder. London WC1H 9SH, UK [email protected] Interpretation This is the fi rst health study of a large and diverse sample of men, women, and child survivors of traffi cking for various forms of exploitation. Violence and unsafe working conditions were common and psychological morbidity was associated with severity of abuse. Survivors of traffi cking need access to health care, especially mental health care.

Funding Anesvad Foundation and International Organization for Migration International Development Fund.

Copyright © Kiss et al. Open Access article distributed under the terms of CC BY-NC-ND.

Introduction vulnerability or of the giving or receiving of payments or Traffi cking of human beings is a gross violation of human benefi ts to achieve the consent of a person having control rights that often involves extreme forms of abuse and over another person, for the purpose of exploitation. exploitation. Recent estimates suggest that more than Exploitation shall include, at minimum, the exploitation of 18 million people are in forced labour as a result of prostitution of others or other forms of sexual exploitation, traffi cking, although estimates are often questioned forced labour or services, slavery or practices similar to because of the hidden nature of traffi cking and diffi culties slavery, servitude or the removal of organs.” in defi ning it.1 Experts generally agree that the crime of human The most commonly used defi nition of human traffi cking traffi cking centres on acts of exploitation and coercion.3 comes from the UN Convention Against Transnational Other more specifi c elements that are commonly present Organized Crime and its Protocol to Prevent, Suppress and in defi nitions of traffi cking and used in treaties, policy Punish Traffi cking in Persons, Especially Women and documents, and operational indicators of traffi cking Children:2 “The recruitment, transportation, transfer, include restricted freedom of movement, abusive living harbouring or receipt of persons by means of threat or use and working conditions, confi scation of documents, debt of force or other forms of coercion, of abduction, of fraud, bondage, withholding of wages for prolonged periods, of deception, of the abuse of power, or of a position of intimidation, and excessive overtime.4–6 www.thelancet.com/lancetgh Vol 3 March 2015 e154 Articles

People who are subjected to these extreme levels of labour in southeast Asia. We describe the health risks exploitation are exposed to many health risks, including and morbidity patterns, including the eff ects of various physical, sexual, and psychological violence, deprivation, forms of violence, occupational risk exposures, and living and severe occupational hazards, which often result conditions, on survivors’ mental health to inform in acute and long-term physical and psychological provision of services and policy. morbidity—sometimes death.7 For those who survive a traffi cking experience, many—if not most—will need Methods medical care for their physical and psychological health Study design and participants needs.8 Worldwide, post-traffi cking services have arisen We carried out this observational cross-sectional study in to care for survivors of these dangerously exploitative Cambodia, Thailand, and Vietnam. We used a two-stage circumstances, which are variously labelled as, among strategy to identify a sample of men, women, and young other terms, traffi cking, forced labour, and slavery. people (age 10–17 years) who used post-traffi cking Very little survey research has been published about services. First, we selected services in each country (six in human traffi cking and health, with most evidence Cambodia, four in Thailand, and fi ve in Vietnam on the coming from studies of women traffi cked for sexual basis of diversity of clientele (eg, age, sex, sector of exploitation, primarily in Europe and south Asia.7 Almost exploitation, country of origin), relationship with country no research has been done of the health of traffi cked teams of the International Organization for Migration, men and boys,9 or people traffi cked into sectors other and agreements with government agencies (eg, support, than sex work, such as agriculture, commercial fi shing, referral, and service arrangements). Several organisations domestic servitude, factory work, and street begging. had specifi c age and sex eligibility criteria (eg, only Although violence and intimidation are commonly women and children, only men). Second, we invited a associated with human traffi cking, little recognition has consecutive sample of individuals within the fi rst 2 weeks been given to the additional occupational and other of admission to these services to participate in face-to-face health risks of traffi cking-related labour conditions.10 interviews. Individuals were excluded if trained case- For example, migrant workers report more accidents and workers deemed individuals too unwell to participate or injuries than do non-migrant workers,11 suggesting that for that participation would cause harm. traffi cked workers (often deemed a subgroup of the larger Clients were referred to these services by various migrant worker population), common occupational sources, including police and immigration services, hazards may be worse, exacerbated by violence and abusive non-governmental and international organisations, and working and living conditions. Traffi cked people are likely government agencies (eg, Cambodia’s Department of to toil in physically arduous jobs and work extensive hours Anti-Traffi cking and Juvenile Protection, Thailand’s with few breaks—conditions associated with high injury Department of Social Development and Welfare, rates.12 Exploited labourers are unlikely to be off ered Vietnam’s Department of Social Evils Prevention). adequate training (in a language they understand) or The study was approved by the ethics committee of personal protective equipment to, for example, use heavy the London School of Hygiene & Tropical Medicine. equipment, work at heights or with harsh chemicals, or to Local ethics approval was granted by the National Ethics do repetitive tasks (eg, bending, lifting).13 Traffi cked Committee for Health Research in Cambodia, by the workers are also likely to work in sectors with few health Hanoi School of Public Health in Vietnam, and by the and safety inspections.14 Despite the evident health risks Ministry of Social Development and Human Security and probable harm associated with human traffi cking, in Thailand. most information about traffi cking comes from media coverage or reporting of individual cases, whereas survey Procedures data—especially for health—are scarce.15 The questionnaire was based on an instrument from a An estimated 56% of forced labourers worldwide live European study of women in post-traffi cking services19 in the Asia and Pacifi c region,1 especially in the Greater and adapted to local contexts and a wider range of labour Mekong subregion,16 which has more than 13 million sectors by the study team. It included questions about migrant workers.1 A small study of Vietnamese migrants socioeconomic background, pre-traffi cking exposures, estimated that 13% are traffi cked.17 Likewise, 17% of living and working conditions during traffi cking, violence, 596 Thai fi sherman reported that they were working health outcomes, and future plans and concerns. against their will or under threat of penalties.18 Given The instrument was translated into Khmer, Thai, that international and internal economic migration are Vietnamese, Burmese, and Lao, and refi ned through common in the region, these proportions suggest that group discussions with International Organization for migrants have a substantial risk of ending up in Migration counter-traffi cking teams, further revised dangerously exploitative circumstances. through pilot-testing, and reviewed after back-translation We report the fi rst large quantitative study of the health into English. Interviews were carried out by social of women, men, and children in post-traffi cking services workers or caseworkers, following intensive 1-week who were exploited and abused in various forms of training by LK in collaboration with the International e155 www.thelancet.com/lancetgh Vol 3 March 2015 Articles

Organization for Migration partners in each country. wanted. We measured participants’ living conditions as Data collection and entry were coordinated by the local living and sleeping in overcrowded rooms, sleeping in International Organization for Migration offi ces, with dangerous conditions, nowhere to sleep or sleeping on oversight by the London School of Hygiene & Tropical the fl oor, poor basic hygiene, inadequate drinking water, Medicine between October, 2011, and May, 2013. The research teams in each country followed a strict ethics protocol based on the WHO Ethical Recom- Men Women Children Total (n=383) (n=288) (n=344) (n=1015) mendations for Interviewing Traffi cked Women.20 Core ethical guidance included measures to ensure that Age (years) participation was voluntary and confi dential, assurance 10–14 ·· ·· 69 (20·1%) 69 (6·8%) that declining participation would not aff ect provision of 15–17 ·· ·· 275 (79·9%) 275 (27·1%) services, avoidance and management of distress, and 18–24 168 (43·9%) 191 (66·3%) ·· 359 (35·4%) off er of options for supported referral for health or other 25–34 151 (39·4%) 52 (18·1%) ·· 203 (20·0%) reported problems. We measured symptoms of anxiety ≥35 64 (16·7%) 45 (15·6%) ·· 109 (10·7%) and depression with the Hopkins Symptoms Checklist Education and post-traumatic stress disorder with the Harvard Primary or less 177 (46·2%) 92 (31·9%) 144 (41·9%) 413 (40·7%) Trauma Questionnaire. 21–23 We used a cutoff of 1·75 for Secondary 111 (29·0%) 97 (33·7%) 151 (43·9%) 359 (35·4%) anxiety, on the basis of studies of users of post-traffi cking Higher 23 (6·0%) 33 (11·5%) 13 (3·8%) 69 (6·8%) service and studies of Cambodian, Laotian, and No formal education 70 (18·3%) 54 (18·8%) 28 (8·1%) 152 (15·0%) Vietnamese refugees.24,25 We excluded item 12 (sexual Data missing 2 (0·5%) 12 (4·2%) 8 (2·3%) 22 (2·2%) interest) from the depression scale because of sensitivity Country of origin in cases of sexual abuse and because participants were Cambodia 228 (59·5%) 38 (13·2%) 40 (11·6%) 306 (30·1%) often residing in shelter situations. Therefore, we used Laos 2 (0·5%) 35 (12·2%) 74 (21·5%) 111 (10·9%) 1·625 as the cutoff for symptoms indicative of depression, Burma 54 (14·1%) 14 (4·9%) 45 (13·1%) 113 (11·1%) departing from the 1·75 cutoff established by Mollica and Thailand 2 (0·5%) 5 (1·7%) 133 (38·7%) 140 (13·8%) 25 colleagues and assuming that each item made a similar Vietnam 97 (25·3%) 196 (68·1%) 51 (14·8%) 344 (33·9%) contribution to the overall score. We used a cutoff of 2·0 Other 0 (0%) 0 (0%) 1 (0·3%) 1 (0·1%) for post-traumatic stress disorder, on the basis of a Country of destination 26 previous study of users of post-traffi cking services. Cambodia 0 (0%) 6 (2·1%) 1 (0·3%) 7 (0·7%) Items on physical and sexual violence were based on a China 99 (25·8%) 182 (63·2%) 49 (14·2%) 330 (32·5%) 27 WHO international study of domestic violence. The Malaysia 30 (7·8%) 22 (7·6%) 0 52 (5·1%) items were supplemented with acts commonly reported Thailand 92 (24·0%) 70 (24·3%) 286 (83·1%) 448 (44·1%) by traffi cking victims to local service providers. On the Vietnam 0 (0%) 2 (0·7%) 1 (0·3%) 3 (0·3%) basis of a categorisation often used in violence studies,28 Indonesia 124 (32·4%) 0 (0%) 5 (1·5%) 129 (12·7%) we classed the following experiences as severe violence: Mauritius 32 (8·4%) 0 (0%) 1 (0·3%) 33 (3·3%) being kicked, dragged, beaten up, tied or chained, choked, South Africa 6 (1·6%) 0 (0%) 0 (0%) 6 (0·6%) or burned; having a dog released to bite or scratch; being Russia 0 (0%) 6 (2·1%) 0 (0%) 6 (0·6%) threatened with a weapon, cut with a knife, shot; and Don’t know 0 (0%) 0 (0%) 1 (0·3%) 1 (0·1%) forced to have sex. Slaps, pushes, and hits were classed as less severe violence. We also asked participants about Sector of exploitation threats against themselves or family members and people Sex work 1 (0·3%) 127 (44·1%) 201 (58·4%) 329 (32·4%) they cared about. Entertainment or karaoke 0 (0%) 6 (2·1%) 23 (6·7%) 29 (2·9%) Items measuring working and living conditions included Animal farming or meat packing 1 (0·3%) 1 (0·3%) 3 (0·9%) 5 (0·5%) excessive working time, restricted freedoms, being cheated Agriculture, farming, or plantation 19 (5·0%) 36 (12·5%) 3 (0·9%) 58 (5·7%) of wages, and precarious living conditions. Excessive Begging 3 (0·8%) 0 (0%) 22 (6·4%) 25 (2·5%) working time was based on the International Labour Car care 2 (0·5%) 0 (0%) 3 (0·9%) 5 (0·5%) Organization’s International Standards on Working Time,29 Domestic worker or cleaner 2 (0·5%) 26 (9·0%) 10 (2·9%) 38 (3·7%) and combined two variables: hours worked per day and Construction 8 (2·1%) 3 (1·0%) 8 (2·3%) 19 (1·9%) hours worked per week. We defi ned non-abusive working Factory 76 (19·8%) 40 (13·9%) 20 (5·8%) 136 (13·4%) time as 8 h or less of work per day or 40 h per week. We Fishing 262 (68·4%) 0 (0%) 13 (3·8%) 275 (27·1%) classed working time of 8–10 h per day or 40–48 h per week Home business 4 (1·0%) 1 (0·4%) 1 (0·3%) 6 (0·6%) as excessive. Extremely excessive working time included Restaurant or hospitality 0 (0%) 0 (0%) 7 (2·0%) 7 (0·7%) more than 10 h per day or more than 48 h per week, or no Street seller or shop 0 (0%) 4 (1·4%) 14 (4·1%) 18 (1·8%) fi xed hours. Wife 0 (0%) 38 (13·2%) 15 (4·4%) 53 (5·2%) We classed restricted freedom as positive when Other 5 (1·3%) 6 (2·1%) 1 (0·3%) 12 (1·2%) participants reported having been locked in a room or (Table 1 continues on next page) never being free to do what they wanted or go where they www.thelancet.com/lancetgh Vol 3 March 2015 e156 Articles

and time in traffi cking). We selected independent variables Men Women Children Total 4,7,31 (n=383) (n=288) (n=344) (n=1015) related to exploitation, deception, and abuse. The epidemiological hypotheses we tested in the regression (Continued from previous page) models (associations between exploitation, violence, and Time in traffi cking situation (months) abuse and mental health symptoms) were defi ned a priori. <1 25 (6·5%) 62 (21·5%) 73 (21·2%) 160 (15·8%) We did the statistical analysis with Stata (version 13). 1–6 138 (36·0%) 130 (45·1%) 198 (57·6%) 466 (45·9%) 7–12 51 (13·3%) 32 (11·1%) 38 (11·1%) 121 (11·9%) Role of the funding source 13–23 47 (12·3%) 21 (7·3%) 7 (2·0%) 75 (7·4%) The funders were invited as observers in the planning ≥24 106 (27·7%) 23 (8·0%) 12 (3·5%) 141 (13·9%) and interpretation meetings, but had no role in data Data missing 16 (4·2%) 20 (6·9%) 16 (4·7%) 52 (5·1%) collection, data analysis, and writing of the report. The Data are n (%). corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to Table 1: Participant characteristics submit for publication.

Results Men (n=383) Women (n=288) Children (n=344) Total (n=1015) Our sample consisted of 1102 participants from 15 post- Physical violence* 188 (49·1%) 118 (41·3%) 82 (23·8%) 388 (38·3%) traffi cking services, of whom 1015 reached work Sexual violence† 5 (1·3%) 125 (43·9%) 74 (21·5%) 204 (20·2%) destinations (appendix). The response rate in all Physical or sexual 188 (49·3%) 171 (60·0%) 122 (35·5%) 481 (47·6%) three countries was more than 98%. Participants reported violence, or both‡ diverse socioeconomic and demographic characteristics, Threats 225 (58·7%) 151 (52·4%) 102 (29·7%) 478 (47·1%) origin and destination countries, sectors of exploitation, Never free to do what 280 (73·1%) 197 (68·6%) 119 (34·6%) 596 (58·8%) wanted§ and conditions experienced during traffi cking (tables 1, 2). Locked in a room 67 (17·5%) 75 (26·0%) 56 (16·3%) 198 (19·5%) Most of the children we interviewed were girls (281 of 344 [82%]). Most participants (146 of 1015 [86%]) were Data are n (%). *Two missing data. †Six missing data. ‡Five missing data. §One missing datum. traffi cked across borders. Individuals traffi cked internally Table 2: Reports of violence and movement restrictions were mainly from Thailand (138 of 144 [95%]). Men were mainly exploited in the fi shing sector and factories. The main destination countries for these men See Online for appendix insuffi cient food (often hungry), no clean clothing, and were Indonesia and China. More than half of men were overexposure to sun or rain. in traffi cking situations for 7 months or more (table 1). We developed the physical health variables on the Women were often exploited in sex work or factories, or basis of an adapted version of the Miller Abuse scale30 were traffi cked as brides. Their main destinations were and medical review of systems (eg, neurological, China and Thailand. Almost half of women were musculoskeletal) used in clinical settings. Participants traffi cked for 1–7 months (table 1). were asked about health problems experienced in More than half of children were traffi cked for sex work, the past 4 weeks and variables were coded as positive with 201 of 281 (72%) of girls forced into sex work. Boys for people who reported severe levels (“extremely” were mainly traffi cked for begging (20 of 63 [32%]), and “quite a lot”). We classed participants as having fi shing (13 of 63 [21%]), factory work (nine of 63 [14%]), had been cheated of wages if they did not receive any and construction work (eight of 63 [13%]). The main cash payments. destination for children was Thailand. Almost all children were in traffi cking situations for less than 12 months. Statistical analysis Almost half of participants experienced physical violence, The analysis includes only individuals who reached the sexual violence, or both, including the majority of adults work destination (which might have entailed same-country (eg, slapped, shoved, or had something thrown that could [internal] or cross-border migration). Although people hurt; pushed or shoved; hit with a fi st or with something often experience violence before reaching a destination, else that could hurt; kicked, dragged, or beaten up; see we focused on health exposures and abusive circumstances table 2). Almost half of men reported physical violence and in the destination location. some reported sexual abuse. Among women, sexual abuse We calculated proportions for all variables and included was much more common and physical violence was 95% CI for prevalence estimates for mental health slightly less common than in men. More than a third of outcomes. We used Cronbach’s α to assess the reliability children reported physical violence, sexual violence, or of the summary scores for the anxiety, depression, and both; just over a fi fth reported sexual violence and almost a post-traumatic stress disorder scales. quarter reported physical violence. We used adjusted logistic regression models to estimate 55 (20%) of 281 girls reported physical violence and the eff ect of traffi cking on mental health outcomes, 73 (26%) of 281 reported sexual violence, whereas 27 (43%) controlling for confounders (age, sector of exploitation, of 63 boys reported physical violence and one (2%) of e157 www.thelancet.com/lancetgh Vol 3 March 2015 Articles

63 reported sexual violence. Overall, 198 (35%) of Men Women Children Total 566 women and girls experienced sexual violence. Threats (n=383) (n=288) (n=344) (n=1015) were reported by almost half of all participants (table 2). Hours worked per day* Extreme restriction of movement—never being free— ≤8 19 (5·0%) 30 (10·8%) 140 (42·3%) 189 (19·1%) was reported by roughly three-quarters of men, two-thirds 8–10 33 (8·7%) 26 (9·4%) 43 (13·0%) 102 (10·3%) of women, and a third of children. 87 (31%) of 284 girls >10 178 (46·8%) 70 (25·2%) 48 (14·5%) 296 (29·9%) and 32 (51%) of 63 boys reported never being free. A high proportion of individuals who were exploited in domestic No fi xed hours 150 (39·5%) 152 (54·7%) 100 (30·2%) 402 (40·7%) work (33 of 38 [87%]), fi shing (221 of 275 [80%]), and as No weekly rest day† 337 (88·7%) 153 (56·0%) 195 (58·7%) 685 (69·5%) wives (38 of 53 [72%]) reported never being free to do what Cheated of wages‡ 288 (75·2%) 231 (80·5%) 114 (33·2%) 633 (62·5%) they wanted or go where they wanted. A fi fth of participants Living conditions reported having been locked in a room (table 2). Living and sleeping in overcrowded 290 (75·7%) 89 (30·9%) 74 (21·5%) 453 (44·6%) rooms 70% of participants reported working 7 days per week, Sleeping in dangerous conditions 109 (28·5%) 12 (4·2%) 23 (6·7%) 144 (14·2%) with almost half of men working 10 h or more (table 3). The Nowhere to sleep or sleeping on the 232 (60·6%) 57 (19·8%) 80 (23·3%) 369 (36·4%) mean number of hours worked per day of people working fl oor every day (n=449) was 13·8 (SD 6·6). Those reporting the Poor basic hygiene 211 (55·1%) 52 (18·1%) 44 (12·8%) 307 (30·3%) longest working hours worked in fi shing (mean 18·8, Inadequate water for drinking 158 (41·3%) 30 (10·4%) 28 (8·1%) 216 (21·3%) SD 5·9) and domestic work (mean 15·2, SD 6·6). Being Insuffi cient food 141 (36·8%) 53 (18·4%) 47 (13·7%) 241 (23·7%) cheated of wages was reported by three-quarters of men, No clean clothing 232 (60·6%) 37 (12·9%) 33 (9·6%) 302 (29·8%) four-fi fths of women, and a third of children (table 3). Overexposure to sun or rain 277 (72·3%) 30 (10·4%) 52 (15·1%) 359 (35·4%) Overall, 690 (68%) of 1015 participants reported at least Recurrent common health problems (self-reported) one bad living condition. A high proportion of men Memory problems 85 (22·2%) 38 (13·2%) 35 (10·2%) 158 (15·6%) reported poor living conditions compared with women and Feeling completely exhausted 92 (24·0%) 49 (17·0%) 45 (13·1%) 186 (18·3%) children (table 3). Poor living conditions were especially Dizzy spells 88 (23·0%) 51 (17·7%) 65 (18·9%) 204 (20·1%) high among fi shermen—eg, 243 (88%) of 275 reported Losing consciousness 8 (2·1%) 2 (0·7%) 2 (0·6%) 12 (1·2%) living and sleeping in overcrowded rooms and 145 (53%) of 275 reported having inadequate drinking water. Headaches 84 (21·9%) 62 (21·5%) 69 (20·1%) 215 (21·2%) 222 (22%) of 1015 participants reported sustaining a Dental problems 55 (14·4%) 13 (4·5%) 29 (8·4%) 97 (9·6%) serious injury at work. Of those reporting a serious injury, Weigh loss 91 (23·8%) 35 (12·2%) 13 (3·8%) 139 (13·7%) 73 (33%) reported being seriously injured a few times and Nausea or indigestion 61 (15·9%) 34 (11·8%) 38 (11·1%) 133 (13·1%) 49 (22%) said they were injured many times. The most Diarrhoea 33 (8·6%) 23 (8·0%) 20 (5·8%) 76 (7·5%) commonly reported injuries were deep or long cuts, skin Perisitent coughing 47 (12·3%) 15 (5·2%) 29 (8·4%) 91 (9·0%) damage, and back or neck injury (table 3). Some Back pain 74 (19·3%) 55 (19·1%) 59 (17·2%) 188 (18·5%) participants reported losing a body part (table 3). Deep cuts Skin problems 55 (14·4%) 17 (5·9%) 55 (16·0%) 127 (12·5%) were most commonly reported by fi shermen (87/128 [68%]). Serious injuries and accidents at work§ 137 (37·5%) 47 (17·2%) 38 (11·1%) 222 (22·6%) Skin damage was frequently reported by factory workers Serious injuries (13/19 [68%]) and street sellers (2/3 [67%]). Back and neck Deep or very long cut 89 (23·2%) 16 (5·6%) 10 (2·9%) 115 (11·3%) injuries were often reported by construction workers Very bad burn¶ 24 (6·3%) 6 (2·1%) 1 (0·3%) 31 (3·1%) (2/5 [40%]) and fi shermen (46/128 [36%]). Only 62 (28%) of Serious head injury¶ 26 (6·8%) 8 (2·8%) 3 (0·9%) 37 (3·7%) 222 participants who were seriously injured reported Back or neck injury 49 (12·8%) 10 (3·5%) 8 (2·3%) 67 (6·6%) receiving medical care for the injury. The most prevalent Skin damage or injury 49 (12·8%) 20 (6·9%) 16 (4·7%) 85 (8·4%) health problems that participants reported experiencing Broken bone 5 (1·3%) 1 (0·3%) 3 (0·9%) 9 (0·9%) quite a lot or extremely were headaches, dizzy spells, back Lost a body part 6 (1·6%) 0 (0%) 1 (0·3%) 7 (0·7%) pain, feeling completely exhausted, and memory problems Eye injury or damage 16 (4·2%) 1 (0·4%) 1 (0·3%) 18 (1·8%) (table 3). A high proportion of men reported weight loss Ear damage 16 (4·2%) 3 (1·0%) 2 (0·6%) 21 (2·1%) and memory problems (table 3). Data are n (%). *26 missing data. †30 missing data. ‡Three missings data. §32 missing data. ¶Two missing data. The scales’ reliabilities were high, with an α of 0·86 for anxiety, 0·89 for depression, and 0·89 for post-traumatic Table 3: Exploitation during traffi cking, common physical health problems, and occupational injuries stress disorder. The mean score for anxiety was 1·74 (SD 0·60), with three quarters of the sample scoring participants reported symptoms of post-traumatic stress 2·1 or more. The overall prevalence of anxiety was 42·8% disorder (table 4). 53 (5%) of 1015 participants reported (95% CI 39·8–45·9; table 4). The mean score for depression having attempted suicide in the month before the interview. was 1·89 (SD 0·64), with more than half of the participants A similar proportion of adult males and females scoring 1·86 or more. Symptoms of depression were reported symptoms of anxiety, with fewer reports from reported by 61·2% (95% CI 58·2–64·2) of participants children (table 4). 30·2% (95% CI 19·9–42·9) of boys and (table 4). The mean score for post-traumatic stress disorder 32·7% (27·5–38·5) of girls reporting symptoms of anxiety. was 1·82 (SD 0·60) and 38·9% (95% CI 36·0–42·0) of More women than men or children reported symptoms www.thelancet.com/lancetgh Vol 3 March 2015 e158 Articles

Men Women Children Total N Prevalence (95% CI) N Prevalence (95% CI) N Prevalence (95% CI) N Prevalence (95% CI) Depression 232 60·7% (5·7–65·5) 191 66·6% (60·9–71·8) 197 57·3% (52·0–62·4) 620 61·2% (58·2–64·2) Anxiety* 185 48·4% (43·4–53·5) 138 48·1% (42·3–53·9) 111 32·3% (27·5–37·4) 434 42·8% (39·8–45·9) PTSD† 177 46·3% (41·4–51·4) 126 43·9% (38·2–49·7) 91 26·5% (22·1–31·5) 394 38·9% (36·0–42·0) Attempted suicide in 16 4·2% (2·6–6·7) 18 6·3% (4·0–9·9) 19 5·5% (3·5–8·5) 53 5·2% (4·0–6·8) the past month†

*Two missing data. †Three missing data.

Table 4: Mental health outcomes

Discussion PTSD Anxiety Depression As the largest survey to date of the health of traffi cking Overtime* survivors, our study confi rms the high levels of various Standard legal working time 1 (reference) 1 (reference) 1 (reference) forms of abuse and serious harm associated with human Excessive 1·45 (0·55–3·83) 1·60 (0·69–3·70) 1·37 (0·62–3·02) traffi cking (panel). Men, women, and children traffi cked Extremely excessive 3·70 (2·02–6·78) 2·09 (1·24–3·52) 1·80 (1·12–2·89) for various forms of forced labour and sexual exploitation Restricted freedom† were highly exposed to physical and psychological abuse, No 1 (reference) 1 (reference) 1 (reference) lived and work in extremely hazardous conditions, and Yes 2·05 (1·49–2·80) 1·86 (1·37–2·53) 1·70 (1·26–2·31) reported serious health problems. This study builds on a Cheated of wages† small body of evidence, primarily on the health of girls No 1 (reference) 1 (reference) 1 (reference) and women traffi cked for sex work, by adding fi ndings Yes 2·01 (1·47–2·76) 1·49 (1·10–2·03) 1·22 (0·89–1·66) about the health needs of men, women, and children Living conditions traffi cked into various labour sectors and off ers unique No bad situations 1 (reference) 1 (reference) 1 (reference) data from the Mekong region.7 At least one bad situation 3·01 (1·96–4·62) 3·08 (2·12–4·47) 2·19 (1·53–3·17) Our fi ndings show that no single profi le of a traffi cked Threats individual exists. Survivors diff ered by age, sex, home No 1 (reference) 1 (reference) 1 (reference) country, and exploitation experiences, suggesting that users Yes 2·70 (2·00–3·64) 2·50 (1·87–3·34) 2·36 (1·75–3·18) of post-traffi cking services are more diverse than only Violence women traffi cked for forced sex work. However, although No violence 1 (reference) 1 (reference) 1 (reference) this diversity suggests that practitioners must treat Less severe 1·21 (0·75–1·95) 1·06 (0·66–1·71) 1·30 (0·81–2·09) individuals and their experiences as unique, the results also More severe 2·02 (1·45–2·82) 1·99 (1·44–2·75) 2·02 (1·45–2·81) suggest common patterns of abuse, occupational risk, and health consequences. These patterns of morbidity off er Data are adjusted odds ratio (95% CI). Models are adjusted by age, sector of exploitation, and time in traffi cking valuable evidence for planning and budgeting for the situation. PTSD=post-traumatic stress disorder. *Defi ned according to the International Labour Organization International Standards on Working Time, with 26 data missing. †One missing datum. ‡Two missing data. health needs of survivors of traffi cking. Our fi ndings confi rm that physical, sexual, and Table 5: Multivariate analysis of factors associated with PTSD, anxiety, and depression psychological abuses are signature features of human traffi cking. Roughly half of participants were physically or associated with depression (table 4). 36·5% (95% CI sexually abused—many suff ering extraordinary forms of 25·3–49·3) of boys and 61·9% (56·0–67·4) of girls scored violence (eg, knife and dog attacks, burning, and choking). positive for depression. Symptoms of post-traumatic Unsurprisingly, most sexual violence was reported by stress disorder were more common in men and women women and girls. than in children (table 4). A higher proportion of girls Our fi ndings also show that restricted freedom is a core (28·1%, 95% CI 23·1–33·7) than boys (19·4%, 11·2–31·5) indicator of traffi cking and a key risk factor for poor reported symptoms of post-traumatic stress disorder. mental health—participants who were severely restricted The multivariate analysis showed strong associations were roughly twice as likely to report symptoms of between abusive and exploitative conditions during post-traumatic stress disorder, anxiety, and depression as traffi cking and poor mental health outcomes (table 5). traffi cked people who were not restricted. Individuals who experienced extremely abusive overtime Our results also showed the relentless days and hours at work, restricted freedom, bad living conditions, that people were made to work. Occupational health threats, or severe violence were more likely to report studies suggest that long hours of working without breaks symptoms of depression, anxiety, and post-traumatic increases the risk of injury33 and exhaustion, which can stress disorder. Having been cheated of wages was have long-term eff ects, such as increased illness and poor signifi cantly associated with anxiety and post-traumatic mental health,34 potentially exacerbated by poor living stress disorder (table 5). conditions35 and unpaid work.36 Our fi ndings show that e159 www.thelancet.com/lancetgh Vol 3 March 2015 Articles

excessive working hours, poor living conditions, and being cheated of wages increased the risk of symptoms of Panel: Research in context post-traumatic stress disorder, anxiety, and depression. Systematic review Although occupational health of traffi cked people has A systematic review of studies of prevalence or risk of violence while traffi cked or any not been well-studied, our study suggests that measure of physical, mental, or sexual health among traffi cked people was done in 2012 work-related health and safety risks must be considered by Oram and colleagues.7 It identifi ed 19 studies, all of traffi cked women and girls, as sector-specifi c—eg, fi shermen suff ered deep cuts from focusing primarily on traffi cking for sexual exploitation. The fi ndings suggested a high sharp knives and dehydration from long hours in sun and prevalence of violence and mental distress, but were limited by methodological 37 at sea, whereas agricultural, animal farm, construction, weaknesses and poor comparability and generalisability. Since 2012, we identifi ed one and domestic workers were repeatedly exposed to study of the health of traffi cked men9 and an analysis of risk factors for mental disorders 38–40 hazardous substances (eg, dust, chemicals, pathogens). among traffi cked women in Moldova.32 Participants described various physical health complaints, but symptoms of poor mental health were most prevalent Interpretation and severe. Symptoms associated with depression, anxiety Our fi ndings suggest that women, men, and children using post-traffi cking services in disorders, and post-traumatic stress disorder seemed to be Cambodia, Thailand, and Vietnam have been exposed to high levels of physical and sexual more common in our sample than in a general population abuse and serious work-related health risks. Women and children are particularly likely to of labour migrants, but similar to those of refugee experience sexual violence. Symptoms of depression, post-traumatic stress disorder, and populations.41 The associations between the intensity of anxiety are prevalent among traffi cking survivors and highly associated with abusive risk exposures and psychological morbidity suggest that working conditions and violence. These results point to the important role of the health the mental health outcomes of our study population sector—particularly mental health professionals—in responding to the urgent and are probably similar to those of other repetitively long-term medical needs of users of post-traffi cking services. Furthermore, the wide trauma-exposed groups. The fi nding that 5% of participants range of labour sectors in which people were abused shows the need for stringent health had attempted suicide in the previous 4 weeks is important and safety standards and regular inspections of low-skilled labour sectors, especially for post-traffi cking health assessments and protection. sectors that are particularly susceptible to human traffi cking. Our study has some limitations. Our sample included only clients of post-traffi cking services, rather than a post-traffi cking service packages. Mental health support general population of traffi cked persons. However, the should be considered an essential component of care. inclusion of many service settings, the comprehensive Intervention research is needed to identify eff ective eligibility criteria, and the large sample size enabled us to forms of psych ological support that can be easily collect data from people of diverse ages, both sexes, and implemented in low-resource settings and in multi- those traffi cked into diff erent sectors. We urge caution lingual, multicultural populations. when comparing subgroups, because some subgroups In view of the wide variety of sectors in which such had small sample sizes. extreme abuses occur, greater government regulation and Data were missing for up to 4% of participants in inspections of low-skilled labour settings—especially those some groups. We believe that this was a result of the known for exploitation—are needed to improve hazardous sensitive nature of the questions being asked or recall working conditions and detect cases of traffi cking. bias in questions asking for details of traffi cking. Exploitation of human beings is age-old. Although it is However, the proportion of missing values was small disheartening to see that human traffi cking exists in and we believe that it had very little eff ect on the results. such proportions in the 21st century, it is nonetheless Finally, the instruments we used to measure mental encouraging that various forms of these violations are health outcomes are not diagnostic and should be increasingly recognised for what they are: modern-day interpreted only as preliminary indicators for emotional slavery. However, alongside global condemnation, there distress and disorder. They have not been validated with needs to be commensurate support for the physical and the study population, but have been used in general psychological health needs of survivors of traffi cking. populations in some of the study countries and in We urge decision makers and donors to invest in post-traffi cking services in Europe.19,42 However, the post-traffi cking health and other services to support the reliability of the scales was high for all three outcomes. recovery of traffi cking survivors. Additionally, the measure of post-traumatic stress Contributors disorder should be interpreted with caution. Because LK was the research coordinator of the study. LK and CZ designed the people were interviewed within 2 weeks of entering the study, analysed and interpreted data, and wrote the report. VN helped to design the study and manage and coordinate regional sites. NSP analysed post-traffi cking service, in some cases we might have and interpreted data and wrote the report. CZ and RB conceived the study. been capturing acute stress disorder rather than PD supported training for fi eldwork and data interpretation. SS, DT, and post-traumatic stress disorder. BD were country coordinators and collected and interpreted data. NP, KS, Our fi ndings show that people who are traffi cked will and VAN collected and interpreted data and revised the report. JK was a country coordinator, helped to design the study, and interpreted data. emerge with a range of health needs and that medical assessment and care to restore people’s physical and Declaration of interests We declare no competing interests. psychological wellbeing should be included in all www.thelancet.com/lancetgh Vol 3 March 2015 e160 Articles

Acknowledgments 23 Mollica R, Caspi-Yarvin Y, Lavelle J, et al. Harvard Trauma This study was funded by Anesvad Foundation and the International Questionnaire (HTQ) Manual: Cambodian, Lao, and Vietnamese Organization for Migration International Development Fund, with Versions. Boston: Harvard Program in Refugee Trauma, 1991. additional support from the Economic and Social Science Research 24 Tsutsumi A, Izutsu T, Poudyal AK, Kato S, Marui E. Mental health Council, UK. of female survivors of human traffi cking in Nepal. Soc Sci Med 2008; 66: 1841–47. References 25 Mollica RF, Wyshak G, de Marneff e D, Khuon F, Lavelle J. 1 ILO. Global Estimate of Forced Labour: Results and Methodology. Indochinese versions of the Hopkins Symptom Checklist-25: Geneva: International Labour Organization, 2012. a screening instrument for the psychiatric care of refugees. 2 Optional Protocol to Prevent, Suppress and Punish Traffi cking in Am J Psychiatry 1987; 144: 497–500. Persons, Especially Women and Children, Supplementing the United 26 Hossain M, Zimmerman C, Abas M, Light M, Watts C. The Nations Convention Against Transnational Organized Crime, GA Res relationship of trauma to mental disorders among traffi cked and 55/25 (2000). http://www.unodc.org/documents/treaties/UNTOC/ sexually exploited girls and women. Am J Public Health 2010; Publications/TOC%20Convention/TOCebook-e.pdf (accessed 100: 2442. Feb 2, 2015). 27 Garcia-Moreno C, Jansen HAFM, Ellsberg M, Heise L, Watts C. 3 Skrivankova K. Between decent work and forced labour: examining WHO multi-country study on women’s health and domestic the continuum of exploitation. Joseph Roundtree Foundation, 2010. violence against women. Geneva: World Health Organization, 2005. http://www.jrf.org.uk/publications/between-decent-work-forced- labour (accessed Sept 17, 2014). 28 Hossain M, Zimmerman C, Kiss L, et al. Men’s and women’s experiences of violence and traumatic events in rural Cote d’Ivoire 4 Gallagher AT. The International Law of Human Traffi cking. before, during and after a period of armed confl ict. BMJ Open 2014; Cambridge: Cambridge University Press, 2010. 4: e003644. 5 Andrees B. Forced labour and human traffi cking: a handbook for 29 ILO. Report III (Part 1B) - General survey of the reports concerning labour inspectors. Geneva: International Labour Organization, 2008. the hours of work (industry) convention, 1919 (No. 1), and the hours 6 UNODC. Abuse of a position of vulnerability and other ‘means’ of work (commerce and offi ces) convention, 1930 (No. 30). In: 93rd within the defi nition of traffi cking in persons. New York: United session. Geneva: International Labour Organization, 2005. Nations Offi ce on Drugs and Crime, 2013. 30 Miller CD, Campbell JC. Reliability and validity of the Miller Abuse 7 Oram S, Stöckl H, Busza J, Howard LM, Zimmerman C. Prevalence Physical Symptom and Injury Scale (MAPSAIS). Chicago, IL: and risk of violence and the physical, mental, and sexual health Midwest Nursing Research Society, 1993. problems associated with human traffi cking: systematic review. 31 ILO. Hard to see, harder to count: survey guidelines to estimate PLoS Med 2012; 9: e1001224. forced labour of adults and children. Geneva: International Labour 8 Borland R, Zimmerman C, eds. Caring for traffi cked persons: Organization, 2011. guidance for health providers. Geneva: International Organization 32 Abas M, Ostrovschi NV, Prince M, Gorceag VI, Trigub C, Oram S. for Migration, 2009. Risk factors for mental disorders in women survivors of human 9 Turner-Moss E, Zimmerman C, Howard LM, Oram S. Labour traffi cking: a historical cohort study. BMC Psychiatry 2013; 13: 204. exploitation and health: a case series of men and women seeking 33 Cooper SP, Burau KE, Frankowski R, et al. A cohort study of post-traffi cking services. J Immigr Minor Health 2013; 16: 1–8. injuries in migrant farm worker families in south Texas. 10 Zimmerman C, Schenker MB. Human traffi cking for forced labour Ann Epidemiol 2006; 16: 313–20. and occupational health. Occup Environ Med 2014; 71: 807–08. 34 Akhavan S, Bildt CO, Franzén EC, Wamala S. Health in relation to 11 Ahonen EQ, Benavides FG, Benach J. Immigrant populations, work unemployment and sick leave among immigrants in Sweden from and health—a systematic literature review. a gender perspective. J Immigr Health 2004; 6: 103–18. Scand J Work Environ Health 2007; 33: 96–104. 35 Schenker MB. A global perspective of migration and occupational 12 Dembe A, Erickson J, Delbos R, Banks S. The impact of overtime and health. Am J Ind Med 2010; 53: 329–37. long work hours on occupational injuries and illnesses: new evidence 36 Devine S. Psychosocial and Mental Health Service Provision for from the United States. Occup Environ Med 2005; 62: 588–97. Survivors of Traffi cking: Baseline Research in the Greater Mekong 13 Cameron L, Lalich N, Bauer S, et al. Occupational health survey of Subregion and Indonesia. Bangkok: International Organization for farm workers by camp health aides. J Agric Saf Health 2006; 12: 139–53. Migration, 2009. 14 ILO. Forced labour and human traffi cking: a handbook for labour 37 Matheson C, Morrison S, Murphy E, Lawrie T, Ritchie L, Bond C. inspectors. Geneva: International Labour Organization, 2008. The health of fi shermen in the catching sector of the fi shing 15 The Lancet. Modern day slavery. Lancet 2007; 369: 800. industry: a gap analysis. Occup Med 2001; 51: 305–11. 16 UNIAP. Mekong Region Country Datasheets: Human Traffi cking 38 Dong XS, Dement J, Welch L, Ringen K. Lifetime risk of 2010. Bangkok: United Nations Inter-Agency Project on Human occupational injuries & illnesses among construction workers. Traffi cking, 2010. Washington, DC: American Public Health Association, 2011. 17 Belanger D. Labor migration and traffi cking among Vietnamese 39 Malhotra R, Arambepola C, Tarun S, de Silva V, Kishore J, Østbye T. migrants in Asia. Ann Am Acad Pol Soc Sci 2014; 653: 87–106. Health issues of female foreign domestic workers: a systematic 18 ILO, Asian Research Center for Migration, Institute of Asian review of the scientifi c and gray literature. Int J Occup Environ Health Studies, Chulalongkorn University. Employment practices and 2013; 19: 261–77. working conditions in Thailand’s fi shing sector. Bangkok: 40 Hribar C. Understanding concentrated animal feeding operations and International Labour Organization, 2013. their impact on communities. Bowling Green: National Association of 19 Zimmerman C, Hossain M, Yun K, Roche B, Morison L, Watts C. Local Boards of Health, 2010. http://www.cdc.gov/nceh/ehs/docs/ Stolen smiles: a summary report on the physical and psychological understanding_cafos_nalboh.pdf (accessed Feb 2, 2015). health consequences of women and adolescents traffi cked in Europe. 41 Lindert J, Ehrenstein OS, Priebe S, Mielck A, Brähler E. Depression The London School of Hygiene & Tropical Medicine, 2006; pp 23. and anxiety in labor migrants and refugees—a systematic review and 20 Zimmerman C, Watts C. WHO Ethical and Safety meta-analysis. Soc Sci Med 2009; 69: 246–57. Recommendations for Interviewing Traffi cked Women. Geneva: 42 Mollica RF, Brooks R, Tor S, Lopes-Cardozo B, Silove D. The World Health Organization, 2003. enduring mental health impact of mass violence: a community 21 Bolton P. Cross-cultural validity and reliability testing of a standard comparison study of Cambodian civilians living in Cambodia and psychiatric assessment instrument without a gold standard. Thailand. Int J Soc Psychiatry 2014; 60: 6–20. J Nerv Ment Dis 2001; 189: 238–42. 22 Lee B, Kaaya SF, Mbwambo JK, Smith-Fawzi MC, Leshabari MT. Detecting depressive disorder with the Hopkins Symptom Checklist-25 in Tanzania. Int J Soc Psychiatry 2008; 54: 7–20.

e161 www.thelancet.com/lancetgh Vol 3 March 2015